A new study in the Annals of Internal Medicine will help reopen the discussion about what extent health insurance decreases mortality and morbidity. What the underlying meaning and importance of these results will long be debated.
Well respected researchers Benjamin Sommers, Sharon Long, and Katherine Baicker evaluated the mortality rates in Massachusetts in the period after the passage of Romneycare, and compared the health results to similar surrounding Northeastern states.
They estimate that overall mortality in Massachusetts declined 2.9 percent relative to control counties between 2007 and 2010. Additionally, death from causes that are ‘amenable’ to health care (such as cancer, heart disease, etc, which potentially can be treated by the health care system) declined 4.5 percent.
This ultimately translated to the saving of 1 life out of every 830 people in the population. As expected, the populations that had the greatest increase in health insurance access showed the greatest improvement.
There has been past debate about whether expansion of health care coverage saves lives. There were mixed results from various studies over the past few decades regarding this subject, but overall, I have never doubted that adequate access to health care should improve health care outcomes.
Jonathan Cohn does point out the obvious quandary: what about the Oregon Medicaid study? If you recall, the Oregon study was a randomized study that showed that those that received Medicaid did not show better health outcomes than those that received no insurance coverage at all. How could this possibly correlate with the findings of this new Massachusetts study?
There are several possibilities that arise. The first is that the improvements in mortality in Massachusetts did not come from the Medicaid population at all, but from those that purchases private insurance. The Massachusetts study does not discriminate between the two populations, so we simply don’t know where the mortality benefits came from, or if they were truly equal across the board, regardless of the type of insurance obtained.
A second possibility (one that researchers Aaron Carroll and Austin Frakt have argued) is that the Oregon study sample size was too small to draw any firm conclusions about demonstrable health effects, and that if the study was large enough, we would have seen these same demonstrable positive effects.
The problem with both of these guesses is that they are central (albeit minor) deficiencies in these two excellent studies, as the authors of both studies themselves readily admit.
My guess (and right now it is only a guess because we have limited evidence at hand) is that health insurance of all types likely decreases morbidity and mortality…even Medicaid. However, I also would guess that the efficiency of those types of insurance in providing adequate care varies greatly; that would correlate with the findings of the Oregon Medicaid study as well.
However, as we go forward, I think there will be a larger issue that we will have to face, one that Austin Frakt mentioned and I seconded on Twitter: what is the cost/benefit ratio of these expenditures?
One thing that the Massachusetts study fails to tell us is if there is a demonstrable difference between Medicaid and other private insurance, as the Oregon study seems to imply. We simply don’t know the answer to that, because the evidence is lacking, other than the Oregon study itself. Why is this important? Currently our data on the ACA expansion of insurance coverage shows that the largest decrease of the number of uninsured likely came from placing people on the rolls of Medicaid; if this is the case, will we see a similar effect as that in Massachusetts, where a higher percentage of people received private insurance?
Secondly, there is a horrible, utilitarian, actuarial question in all this: How much does the Massachusetts study suggest is the cost of saving a single life? As stated above, the improvement would theoretically save 1 life out of every 830 people in the population. If the average cost of health insurance for every person is approximately $4,000, then the cost of saving that single life, per year, is $3,320,000. Is that worth it? Don’t ask me; that is a question for people in a much higher pay grade than mine.
Let me provide another actuarial puzzle to think about. There have a myriad of articles regarding the utility and cost efficiency of mammography, with many of the opinions today saying mammography is not worth the cost, both in dollars and in unnecessary procedures and testing. However, the cost of saving a single life through mammography is approximately similar or even less than the cost of saving a single life through the Massachusetts insurance expansion, if you take this study at face value. Those calculations can vary depending on which mammography study you utilize, I fully admit. However, the controversy over the actual numbers and specifics really is not the point here; the point is, the cost of saving a single life is expensive, both using mammography, as well as in the wider health care arena.
Again, putting a dollar value on ‘a life saved’ is something I don’t want to be personally responsible for…but if we are going to have the debate, someone will have to take a stand on the question.
On a political note, I don’t think this really moves the needle in either direction. Conservative health policy experts such as Avik Roy, Philip Klein, Scott Gottlieb and others all agree that we have to find ways to expand universal access to health care for all; fundamentally, in the many discussions we have had, what is clear is that we all believe access to HEALTH CARE decreases morbidity and mortality. The method of obtaining that access remains in dispute.
So the problem here is not that health care makes people healthier. Dear God, as physician myself, I hope that is the case…or I should go and find another profession.
What all of these experts, myself included, are arguing is fundamentally a different question: what is the most cost efficient and patient-centric method of providing that access? The central question is how to use our limited dollars in the most efficient way to provide the maximal effect of those dollars.
Is expanding Medicaid the most efficient use of those dollars? Are we better off moving those patients into the private market? Are we better off moving to single payer? Would a more patient-centric approach provide better or worse outcomes? All of these questions largely remain unanswered.
I do think that this Massachusetts study, along with the data from the Oregon study, significantly expands our understanding of how health insurance, as it currently exists in the U.S., affects mortality and other factors. Neither study is perfect, and each has its limitations. However, both are well done, and have been conducted brilliantly by well respected researchers; the data is sound. The question really becomes what is the significance of that data, how does it effect the reality of health insurance as it exists today, and furthermore, how do we maximize those effects for the least amount of cost?